0026-895X/07/7101-250 –258$20.00
MOLECULAR PHARMACOLOGY
Copyright © 2007 The American Society for Pharmacology and Experimental Therapeutics
Mol Pharmacol 71:250–258, 2007
Vol. 71, No. 1
29579/3163397
Printed in U.S.A.
Cellular Retinoic Acid Bioavailability Determines Epithelial
Integrity: Role of Retinoic Acid Receptor ␣ Agonists in Colitis
Makoto Osanai, Nami Nishikiori, Masaki Murata, Hideki Chiba, Takashi Kojima,
and Norimasa Sawada
Departments of Pathology (M.O., M.M., H.C., T.K., N.S.) and Ophthalmology (N.N.), Sapporo Medical University
School of Medicine, Sapporo, Japan
Received August 8, 2006; accepted October 11, 2006
Tight junction (TJ) is the apical-most intercellular structure in epithelial cells, playing a role in cell-cell adhesion,
polarity, and the permeability barrier to paracellular transport of solutes across the cells (Tsukita et al., 2001; Matter et
al., 2005). The TJ creates a semipermeable barrier, separating different organ compartments. Whereas the TJ is composed of a large number of protein components, including
membrane proteins such as occludin and cytoplasmic scaffolding proteins such as zonula occludens-1 (ZO-1), it is now
well accepted that the claudin family, which has been shown
to contain more than 20 members, is the main constituent of
the TJ, with tissue-specific distribution of its expression. The
epithelial barrier is critical to maintain tissue homeostasis;
however, only a few agents that can enhance or protect epithelial barrier function have been proposed (Prosser et al.,
2004; Howe et al., 2005).
This study was supported by grant (to N.S.) from the Ministry of Education,
Culture, Sports, Science, and Technology of Japan.
Article, publication date, and citation information can be found at
http://molpharm.aspetjournals.org.
doi:10.1124/mol.106.029579.
tracer molecules, inulin and mannitol. This RA-mediated enhancement of barrier function is potentially associated with the
increased expression of TJ-associated genes such as occludin,
claudin-1, claudin-4, and zonula occludens-1. We also found
that RAR␣ is a preferential regulator of the epithelial barrier in
vitro. Studies of murine experimental colitis, which is characterized by increased gut permeability, reveal that RAR␣ stimulation significantly attenuates the loss of the epithelial barrier
during colitis in vivo. Our results suggest that cellular RA bioavailability determines the epithelial integrity, because it is a
critical regulator for barrier protection during mucosal injuries.
Retinoic acid (RA) is a biologically active regulator that has
a broad range of functions involving cell differentiation, proliferation, and apoptosis in various cell types (Durston et al.,
1989; Osanai and Petkovich, 2005). The effects of endogenous
RAs are achieved primarily by two types of RA isomers (alltrans RA [atRA] and 9-cis RA [9cRA]), which are synthesized
from the retinalaldehyde precursor, a derivative of vitamin
A, and are mediated by two classes of nuclear receptors,
retinoic acid receptors (RAR␣, ␤, and ␥) and their heterodimeric counterparts, retinoid X receptors (RXR␣, ␤, and
␥) (Kastner et al., 1995). RARs are activated both by atRA
and 9cRA, but RXRs are selectively activated by 9cRA. Accumulated evidence supports the possibility that atRA is an
obligatory component in the differentiation of epithelial cells
that leads to the establishment of epithelial integrity,
whereas only a very limited number of reports have documented the functions of atRA in TJ formation. One example
is based on our observation that atRA plays an important
role in the formation of functional TJs in F9 embryonal
carcinoma cells through the involvement of specific RAR/
ABBREVIATIONS: TJ, tight junction; atRA, all-trans retinoic acid; RA, retinoic acid; RAR, retinoic acid receptor; RXR, retinoid X receptor; ZO-1,
zonula occludens-1; 9cRA, 9-cis retinoic acid; MDCK, Madin-Darby canine kidney; TER, transepithelial electrical resistance; RT-PCR, reverse
transcription-polymerase chain reaction; TNBS, 2,4,6-trinitrobenzene sulfonic acid; DMSO, dimethyl sulfoxide; Am80, 4[(5,6,7,8-tetrahydro5,5,8,8-tetramethyl-2-naphthalenyl)carbamoyl]benzoic acid; Am580, 4(5,6,7,8-tetrahydro-5,5,8,8-tetramethyl-2-naphtamido)benzoic acid; FITC,
fluorescein isothiocyanate; PPAR, peroxisome proliferator-activated receptor; FBS, fetal bovine serum.
250
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
ABSTRACT
The epithelial barrier is determined primarily by intercellular
tight junctions (TJs). We have demonstrated previously that
all-trans retinoic acid (atRA) plays an important role in forming
functional TJs through a specific retinoic acid receptor (RAR)/
retinoid X receptor heterodimer in epithelial cells. However, the
physiological relevance of retinoic acids (RAs) in maintaining
the epithelial integrity remains to be examined. Here, we show
that several types of RA, including atRA, promote the barrier
function of epithelial TJs. Conversely, RA depletion in the cells
by overexpressing CYP26s, cytochrome P450 enzymes specifically involved in the metabolic inactivation of RAs, induces an
increase of permeability as measured by two differently sized
Role of Retinoic Acid in Epithelial Barrier
Materials and Methods
Cell Line and Culture. Madin-Darby canine kidney (MDCK)
cells were maintained in DMEM (Invitrogen, Tokyo, Japan) supplemented with 10% FBS (Sanko Jyunyaku, Tokyo, Japan), 10 mM
HEPES, 100 U/ml penicillin, and 100 ␮g/ml streptomycin at 37°C in
a humidified 5% CO2 atmosphere.
Expression Vector and Transfection. Full-length human
CYP26A1 cDNA (the coding region is from nucleotide 26 to 1519
[CYP26A126–1519], GenBank accession number NM_000783) was
amplified by reverse transcription (RT)-PCR from 1 ␮g of total
RNA extracted from atRA-treated MCF7 cells for 24 h (White et al.,
1997). We also amplified truncated forms of CYP26A126–1519,
CYP26A126–1338, CYP26A126–976, and CYP26A1396–1519 using sitespecific primers. After subcloning into a TA-cloning vector (pCRII)
using a TA cloning kit (Invitrogen), the digested approximately 1.5-,
1.3-, 0.9-, and 1.1-kilobase EcoRI fragments were ligated into the
response plasmid pcDNA3.1(⫺) (Invitrogen). A CYP26C1 constitutive expression vector was also made by the insertion of full-length
CYP26C1 cDNA in frame into pcDNA3.1(⫺). We verified the integrity of final constructs by direct sequence analysis. All primers used
to make these constructs are available by request.
Five micrograms of each plasmid was transfected into MDCK cells
using FuGENE 6 reagent (Roche Diagnostics, Tokyo, Japan). G418
(Sigma, Tokyo, Japan)-resistant clones were expanded as a monoclonal population. Two weeks later, cell lines were selected to examine
the strong expression of introduced gene by Northern blot analysis,
and metabolic activities in each cell line were determined by measuring the metabolism of radiolabeled RA ([3H]atRA and [ 3H]9cRA;
PerkinElmer Life and Analytical Sciences, Yokohama, Japan)
(White et al., 1997). Because our preliminary experiments showed
that observed phenotypes of the cells in the event of CYP26A1
overexpression were similar but not the same between at least three
different independent clones, these cells were equally mixed to establish stably transfected cell lines to avoid possible clonal variation.
Assessment of Epithelial Permeability. We measured the
transepithelial electrical resistance (TER) and paracellular flux. For
TER analysis, cells were grown to confluence on Transwell membranes (pore size, 0.4 ␮m; Corning Life Sciences, Acton, MA) in the
presence or absence of 100 nM atRA, 100 nM 9cRA, or 100 nM
synthetic retinoids as a default concentration. TER was measured
using an EVOM voltmeter with ENDOHM-12 (World Precision Instruments, Sarasota, FL). For calculation of resistance, the background TER of a blank filter was subtracted from the measured
values, and each value was normalized by the area of the monolayer
expressed in standard units of ohms per square centimeter. Paracellular tracer flux was measured with the radiolabeled molecules having two different molecular weights by using [14C]inulin (molecular
mass, 5 kDa) and [14C]mannitol (molecular mass, 182 Da) (GE
Healthcare, Tokyo, Japan). An epithelial monolayer treated with or
without RA was exposed to [14C]inulin or [14C]mannitol in the apical
compartments to adjust the radioactivity using 5 ⫻ 105 or 1 ⫻ 105
cpm/well, respectively. Samples were collected from the basolateral
compartment (outer chamber) in a time-dependent manner, and the
radioactivity of [14C] was counted by scintillation counter (Beckman
Coulter, Fullerton, CA). The paracellular flux was normalized by the
value measured in blank Transwell filters. In some experiments,
ketoconazole (0.5–10 ␮M, Sigma-Aldrich) was added 4 h before RA
stimulation and then maintained throughout the experiments.
RT-PCR and Southern Blotting. After extraction of total RNA
using TRIzol (Gibco BRL) reagent, total RNA (1 ␮g) was reversetranscribed using MuLV reverse transcriptase (Applied Biosystems,
Foster City, CA). For analysis of gene expression, the gene of interest
was amplified from dilutions of cDNA using specific primers for
between 15 and 40 cycles to define the optimal conditions for linearity to permit quantitative analysis of signal length. As a positive
control, and to confirm that each RNA sample could yield equal
amounts of product after reverse transcription reactions, amplification was also done using glyceraldehyde-3-phosphate dehydrogenase
primers. Agarose gels were then transferred to Hybond-N⫹ membranes (Amersham Biosciences), and the blots were probed with the
corresponding random-primed 32P-labeled cDNA (Nippon Gene, Tokyo, Japan) with a specific activity of approximately 1 ⫻ 109 cpm/␮g.
For the densitometric analysis, signals in the blot were quantitated
using Scion Image 1.62 (Scion Corporation, Frederick, MD).
Western Blot Analysis. Whole lysates (20 ␮g) extracted from the
cells were run on 12% polyacrylamide gels containing SDS and
electroblotted onto nitrocellulose filters. After blocking with 5% nonfat dry milk in phosphate-buffered saline, filters were immunoblotted with antibodies against occludin (Zymed Laboratories, San Francisco, CA), claudin-1 (Ishizaki et al., 2003), claudin-4 (Ishizaki et al.,
2003), ZO-1 (Santa Cruz Biotechnology, Santa Cruz, CA), and ␤-actin (Santa Cruz Biotechnology) protein. The filters were washed
extensively, reacted with corresponding peroxidase-labeled secondary antibodies, and again washed; finally, the immunoreactions were
visualized by using an enhanced chemiluminescence system (Amersham Biosciences). Equal loading was demonstrated by ␤-actin expression.
Experimental Murine Colitis. 2,4,6-Trinitrobenzene sulfonic
acid (TNBS; Sigma) colitis was induced as described previously (Morris et al., 1989). In brief, mice were sensitized by cutaneous administration of 1% TNBS on day 1 followed by intrarectal administration
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
RXR heterodimer-mediated transcription machinery (Kubota
et al., 2001). On the other hand, loss of expression of TJassociated molecules such as occludin has been correlated
with tumor development in carcinogenesis, and decreased
and/or impaired TJ formation has been reported for various
types of cancer (Tobioka et al., 2004; Osanai et al., 2006). It
is interesting that disruptions of retinoid signaling through
mutations in RARs and RXRs have been shown to be associated with the disorganization of epithelial architecture such
as that observed in cancer tissues (de Thé, 1996). This evidence prompted us to examine the possible hypothesis that
cellular RA bioavailability may play an important role in the
establishment of functional TJs in epithelial cells.
Cellular RA bioavailability is regulated by the vitamin A
nutritional status and the coordinated balance between RA
synthesis and catabolism. RA metabolizing cytochrome P450
enzymes (CYP26s) are specifically involved in the metabolic
inactivation of RA (White et al., 1997; Abu-Abed et al., 2001;
Taimi et al., 2004; Osanai and Petkovich, 2005). A member of
this family, CYP26A1, has been shown to be responsible for
the catabolism of atRA, and CYP26C1 efficiently converts
both atRA and 9cRA to biologically inactive polar metabolites. The expression of CYP26s plays a critical role in embryogenesis by restricting exposure to inappropriate concentrations of RA (Abu-Abed et al., 2001), and a state of atRA
deficiency induced by enhanced expression of CYP26A1 is an
important contributing factor for apoptotic resistance in certain types of cancer (Osanai and Petkovich, 2005). Although
the functional roles of CYP26s remain to be studied, RAcatabolizing enzymes play an important role on limiting the
bioavailability of RA to cells and efficiently caused an RAdepleted state in given cells. By overexpressing CYP26s, we
generated RA depletion in cultured cells to examine whether
RA was an important determinant for the epithelial barrier.
Here, we show that RAs, which stimulate RAR␣, preferentially promote the epithelial barrier function of TJs, which is
sufficient to result in attenuated loss of the barrier during
experimental colitis in vivo, suggesting that RA bioavailability in the cell regulates the epithelial integrity.
251
252
Osanai et al.
Results
atRA Regulates Barrier Function of TJ. Numerous
reports have revealed that MDCK cells are an excellent tool
for studying epithelial permeability because they have
higher TER, abundantly express TJ-associated genes such as
claudin-1, claudin-4, and occludin, and bear well-developed
TJ strands (Furuse et al., 2001). To deplete the cells of atRA,
we established an MDCK cell line, constitutively overexpressing CYP26A1, which is specifically involved in the metabolic inactivation of atRA, designated MDCKCYP26A1. Cells
transfected with an empty vector were used as a control
(MDCKVec). We first examined the effects of atRA on MDCK
cells. Treatment with atRA significantly increased TER in
MDCKVec cells in a time-dependent manner; however,
MDCKCYP26A1 cells showed basal levels of TER after atRA
treatment in the range of 10 nM to 1 ␮M for 3 days (Fig. 1A).
TER reached the maximum in control cells after 5 days in the
presence of ⬃1 ␮M atRA (Fig. 1A). Unexpectedly, higher
concentrations of atRA (⬎1 ␮M) increased TER even in
MDCKCYP26A1 cells, suggesting that CYP26A1 enzymatic activity could not completely catabolize atRA contained in the
media. We next measured the paracellular flux in both cell
lines (Fig. 1B). Treatment with atRA also affected paracellular flux, significantly inhibiting the flux of inulin (5 kDa) but
not of mannitol (182 Da). Breakdown of atRA mediated by
CYP26A1 increased the permeability as measured both by
inulin and mannitol. This observation suggested that a pharmacological dose of atRA regulated paracellular transport of
solutes with high molecular weight but that depletion of a
physiological dose of atRA induced a leaky epithelial monolayer even for solutes with small molecular weight.
Direct Relationship Between atRA Bioavailability
and Epithelial Barrier Function. Ketoconazole, a broadspectrum inhibitor of CYP26s, was next used to establish
whether the effect of CYP26A1 on cells was mediated by the
metabolism of atRA (Osanai and Petkovich, 2005). Ketocon-
azole alone did not significantly affect the values of TER and
flux in MDCKVec cells, whereas treatment of MDCKCYP26A1
cells with ketoconazole partially increased the TER (Fig. 2A)
and conversely decreased the permeability to inulin and
mannitol (Fig. 2B). Ketoconazole treatment did not fully restore the TJ function in MDCKCYP26A1 cells to the level seen
in control cells, suggesting that the period of time required
for clonal expansion in RA-free conditions (due to CYP26A1
overexpression) may have limited the ability of these cells to
respond to various stimuli presented in the media.
To confirm the functional relationship between the atRA
bioavailability and barrier function of epithelial cells, we
made various deletion mutants of CYP26A1 (Fig. 2C) with
different catabolic activities of atRA (Fig. 2D). It is interesting that the cells transfected with deletion mutants of
CYP26A1 showed significant abrogation in decreasing TER
(Fig. 2E) and increasing paracellular flux (Fig. 2F) mediated
by metabolic inactivation of atRA contained in the media,
which was clearly associated with the catabolic activities of
atRA. These data support our novel concept that atRA bioavailability in the cells is an important contributing factor for
determining the epithelial barrier of TJs.
RAR␣ Is an Alternative Regulator to Barrier Function. To use the substrate specificities of CYP26A1 and C1,
we next investigated the role of 9cRA in barrier function.
9cRA could enhance the barrier function, and the cells with
forced expression of CYP26C1 showed significant inhibition
of the increasing effects of TER mediated by both atRA and
9cRA (Fig. 3A). Pretreatment with ketoconazole partially
abrogated this CYP26C1-mediated effect (Fig. 3B). TER recovered ⬃59% in CYP26A1-expressing cells (Fig. 2A) versus
⬃30% in CYP26C1-expressing cells (Fig. 3B) after treatment
with 5 ␮M ketoconazole compared with the cells without
ketoconazole treatment (p ⬍ 0.05), which is consistent with
Fig. 1. Barrier function of TJ is regulated by atRA. A, TER was examined
after 3 and 5 days in the presence or absence of various concentrations of
atRA in MDCKVec (Vec) and MDCKCYP26A1 cells (A1). B, paracellular flux
was measured in a time-dependent manner in the presence (open symbol
and broken line) or absence (filled symbol and solid line) of 100 nM atRA
for 3 days in MDCKVec (squares) and MDCKCYP26A1 cells (triangles),
using two different radiolabeled molecules as tracers. ⴱ, p ⬍ 0.05 versus
cells without atRA treatment.
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
of 5 ␮l/g body weight of 2.5% TNBS solution on day 14. The control
animals received a corresponding volume of vehicle (50% ethanol)
alone. Some animals continued receiving intraperitoneal administrations of a vehicle [dimethyl sulfoxide (DMSO)], atRA (45 mg/m2),
or Am580 (3.75 mg/kg) every 2 days for 1 week. Animals received
these agents as a single initial dose when TNBS was instilled intrarectally under general anesthesia. After treatment, intestinal permeability was assessed using mice that received oral gavage of FITCdextran (4 kDa, 40 mg/kg; Sigma) 24 h before sacrificing animals.
Upon death, blood was collected by cardiac puncture, and the serum
was immediately analyzed for FITC-derived fluorescence (Wallac
1420; PerkinElmer Wallac, Gaithersburg, MD). For all mice, colon
length, which was defined as the distance between the most distal
portion of the cecum and most terminal aspect of the rectum, was
measured, and histological evaluation was done to assess the histological grading of colitis (Dieleman et al., 1998). The maintenance
and handling of animals were carried out using protocols approved
by the Animal Care Facility of Sapporo Medical University.
Assessment of Bacterial Translocation. Spleens from mice
were sonicated in phosphate-buffered saline, and small aliquots (50
␮l) of tissue homogenates were inoculated onto the bacterial agar
plates. Plates were incubated for at least 48 h at 37°C.
Statistical Analysis. All data represent the mean ⫾ S.D. of at
least three independent experiments, each in triplicate wells. Statistical differences were analyzed using the Mann-Whitney U test
and were considered statistically significant when p ⬍ 0.05.
Role of Retinoic Acid in Epithelial Barrier
(Fig. 3, D and E, and data not shown). Although we could not
exclude the participation of the definite RXR subtypes responsible for the modulation of barrier function due to the
lack of specific agonists or antagonists for RXRs, the RAR␣
agonists were, at lease in part, clearly involved in the regulation of barrier function.
We further examined whether atRA attenuated the epithelial damages induced by various stimuli, such as hyperthermia and oxidative stress mediated by H2O2. We exposed
these insults to the epithelial monolayer and showed that
atRA was partially but significantly able to attenuate the
disruption of barrier properties of MDCK monolayer induced
by these stimuli (Fig. 4, A and B). We observed similar effect
on epithelial monolayer by using Am580 (data not shown),
suggesting a possible feasibility of RAR␣-mediated therapy
for diseases that are characterized by epithelial hyperpermeability.
TJ-Associated Genes Are Regulated by Cellular
atRA Bioavailability. We next examined whether changes
Fig. 2. Cellular RA bioavailability regulates epithelial barrier function. A and B, effects of ketoconazole on TER (A) and paracellular flux (B) in
MDCKVec (Vec) and MDCKCYP26A1 (A1) cells after culture with or without 100 nM atRA for 3 days. C, four deletion mutants of CYP26A1 vector. D,
metabolic activities of MDCK cells transfected with each CYP26A1 construct with various deletions using [3H]atRA as a substrate. CYP26A1-mediated
catabolic activity converted radioactive atRA to water-soluble metabolites and was suppressed by pretreatment with ketoconazole. E and F, TER (E)
and paracellular permeability (F) was examined in MDCK cells transfected with each deletion mutant of CYP26A1 treated with (⫹) or without (⫺)
100 nM atRA for 3 days. In measuring TER, before treatment with atRA, the cells were changed to a medium supplemented with 10% charcoal-dextran
FBS (CD-FBS), allowing us to exclude the possible effects of physiological concentrations of atRA contained in the medium. The samples for
paracellular flux were collected after 24 h in the presence of [14C]inulin or [14C]mannitol (B and F). ⴱ, p ⬍ 0.05 versus cells without ketoconazole (A,
B, and D) or atRA (E and F) treatment.
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
the previous observation that CYP26C1 was less sensitive
than CYP26A1 to the inhibitory effects of ketoconazole
(Taimi et al., 2004). CYP26A1 overexpression in MDCK cells
did not have any effect on the 9cRA-mediated increase of
TER (Fig. 3C), suggesting that RXRs activated by 9cRA were
also involved in the regulation of the epithelial barrier.
We next used various types of synthetic retinoids to specifically stimulate and inhibit RAR and RXR subtypes to
confirm the involvement of RARs and RXRs in the regulation
of barrier function. RAR and RXR antagonists partially abrogated atRA- and 9cRA-mediated changes of permeability
(Fig. 3, D and E), suggesting that both RARs and RXRs were
involved in the regulation of epithelial permeability. The
synthetic compounds Am80 and Am580, which are resistant
to metabolism by CYP26s, were reported to stimulate RAR␣
selectively as efficiently as atRA (Luu et al., 2001). In the
dose range from 2.5 to 100 nM, a clear dose-dependence of
enhancement was found even in the CYP26A1-overexpressing MDCK cells that were treated with Am80 and Am580
253
254
Osanai et al.
Treatment with small interfering RNA targeted to occludin did not change the epithelial permeability in MDCK
monolayer (data not shown), which was consistent with the
evidence that the claudin family is the main constituent for
TJ, rather than occludin (Tsukita et al., 2001; Matter et al.,
2005). However, it is agreeable that increased expression of
TJ-associated genes potentially associated with the RA-mediated enhancement of barrier function, because of our previous study demonstrating the direct mechanistic link between atRA-mediated inductions of TJ-associated genes and
the formation of functional TJs in F9 cells (Kubota et al.,
2001).
RA-Mediated Enhancement of Epithelial Barrier
Protects Mice from Colitis. Because RAR␣ is an important
determinant for epithelial barrier in vitro, we used TNBSinduced murine experimental colitis to examine the clinical
relevance of RAR␣ agonists. We first measured the body
weight, because weight loss is a reliable method to determine
the severity of TNBS colitis (Morris et al., 1989). Control
mice lost body weight rapidly and failed to regain the weight
during the course of the 7-day experiment, whereas atRA-
Fig. 3. Preferential interaction between RAR␣ and epithelial barrier function. A to C, TER was examined after treatment with 100 nM atRA or 100
nM 9cRA for 3 days, or various concentrations of ketoconazole in MDCKVec (Vec), MDCKCYP26A1 (A1), and MDCKCYP26C1 (C1) cells. D and E, effects
of various synthetic retinoids on TER (D) and paracellular flux (E) in MDCKVec (Vec) and MDCKCYP26A1 (A1) cells after culture with or without 100
nM synthetic retinoids for 3 days. The samples for paracellular flux were collected after 24 h in the presence of tracers. Lane 1, vehicle (DMSO); lane
2, atRA; lane 3, 9cRA; lane 4, Am80 (RAR␣ agonist); lane 5, Am580 (RAR␣ agonist); lane 6, atRA⫹LE135 (RAR␣⫹␤ antagonist); lane 7, atRA⫹HX531
(pan-antagonist of RARs and RXRs); lane 8, 9cRA⫹HX531; lane 9, 9cRA⫹PA452 (RXR pan-antagonist). ⴱ, p ⬍ 0.05 versus cells without RA treatment.
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
in barrier function regulated by RA bioavailability in the cell
were accompanied by gene-expression alterations of TJ-associated genes. We first examined the quantitativeness of our
method and found that the threshold cycling parameter for
saturation of band intensity showed a clear inverse proportion to the template (Fig. 5A), indicating that our assay was
sensitive to examine the quantification of mRNA expression.
Due to the limited availability of the canine cDNA sequence
of TJ-associated genes, we analyzed genes known to be
cloned such as occludin, claudin-1, claudin-4, and ZO-1 (Fig.
5B). Treatment with atRA significantly up-regulated these
genes in a time-dependent manner, and CYP26A1-mediated
depletion of atRA partially abrogated this induction. Although the basal levels of gene expression were different
between MDCKVec and MDCKCYP26A1 cells, there was evidence that normal FBS contained a physiological dose of
atRA, which was sufficient for gene-expression of the TJassociated genes and functions of TJ (Fig. 2E). In addition,
the TJ protein kinetics in the event of atRA treatment was
consistent with the gene-expression alterations of TJ-associated genes (Fig. 5C).
Role of Retinoic Acid in Epithelial Barrier
Discussion
Here, we demonstrated that several types of RA, including
RAR␣ stimulants, promoted the epithelial barrier function in
vitro and could reduce the breakdown of the gastrointestinal
barrier in vivo. In addition, metabolic depletion of RAs in the
Fig. 4. atRA attenuates the epithelial hyperpermeability induced by
hyperthermia and H2O2. A and B, effects of atRA on TER (A) and paracellular flux (B) in MDCKVec (Vec) and MDCKCYP26A1 (A1) cells after
culture with (⫹) or without (⫺) 100 nM atRA for 3 days. These cells were
also treated with cell stressors including heat shock at 42°C for 1 h and
10 ␮M H2O2 for 24 h. ⴱ, p ⬍ 0.05 versus cells without cell stressor.
cells was clearly proportional to the epithelial barrier dysfunction, suggesting that the cellular RA bioavailability regulates the epithelial barrier accompanied by gene-expression
alterations of TJ-associated genes. Whereas the establishment and regulation of epithelial permeability is the end
result of a cascade of events triggered by cell-cell interactions, our present study provides significant evidence that
cellular RA bioavailability is an important contributing factor for determining the epithelial integrity.
This study focused on mechanisms of the mucosal barrier
protection during intestinal inflammation. There seems to be
a logical gap between in vitro experiments carried out with
MDCK cells and the in vivo study on the experimental colitis
in mice. Accumulated evidence has demonstrated that
MDCK cell represents an excellent model to study general
properties of epithelial barriers; however, there is still a
point of controversy whether the MDCK data are applicable
to the intestinal epithelium. Our preliminary experiment
suggested that the intestinal cell lines such as T84 did not
show significant increase of barrier function after the treatment with atRA. Although we cannot explain this specific
underlying mechanism(s), one possibility is based on the
cellular property showing that T84 cells abundantly express
endogenous CYP26A1 mRNA, which is highly inducible by
the treatment with atRA (our unpublished observation). Endogenous CYP26A1 activity in T84 cells enables us to show
the evidence of whether the exogenously applied atRA would
modulate the epithelial permeability. In addition, we should
note here that the commercially available intestinal cell lines
are originated from cancer tissues that have lost the physiological intestinal barrier function in the carcinogenesis.
This is also supported by the fact that the cancer cells can
hardly be said to faithfully mimic normal biology, in addition
to the fact that their genomes are notoriously diverse and
poorly characterized. One would thus be likely to accept that
MDCK cells were appropriate for this study rather than
cancer-derived intestinal cells.
Intestinal mucosa has crucial functions for regulating intestinal homeostasis by strictly separating the subepithelial
compartment from potentially noxious luminal compounds.
Intestinal barrier dysfunction initiated by various etiologies
is a main contributing factor in several pathological conditions involving the gastrointestinal tract. Increased TJ permeability provides a major site for both infection and establishment of inflammatory responses in the gut. Bacterial
translocation is, for example, the passage of viable bacteria
and nonviable bacteria such as endotoxins and/or bacterial
DNA are believed to be made through the intestinal lumen to
extraintestinal sites via a paracellular pathway between the
epithelial cells, even before massive disruption of the epithelial TJ-based barrier (Van Leeuwen et al., 1994). Our present
study with TNBS colitis revealed a potentially central role for
RA as an endogenous regulator of inflammatory colitis, providing a significant insight into the molecular details linking
retinoid action to the pathology of diseases with gut hyperpermeability.
Given the complexity of the cell signaling events regulating
epithelial permeability (Gordon et al., 2005; Prasad et al.,
2005), it is not surprising that RAs modulate various sets of
genes, including occludin and claudins. Whereas a previous
report has demonstrated that the human occludin promoter
is down-regulated by inflammatory cytokines such as tumor
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
and Am580-treated animals did not show significant weight
loss (Fig. 6A). Likewise, the length of the colon, which shortens as a result of more severe inflammation, was significantly different between the control and RA-treated groups,
in which animals treated with the vehicle showed significant
shortening of the colon, whereas those treated with atRA and
Am580 did not display severe shortening (Fig. 6B). In addition, as assessed by 4-kDa FITC-dextran, treatment with
atRA and Am580 significantly abrogated the increase in intestinal permeability observed in mice with vehicle treatment (Fig. 6C). Furthermore, bacterial translocation, a measure of intestinal barrier integrity, was found in the spleens
of TNBS-treated animals and was correlated with the permeability of FITC-dextran (Fig. 6D). Am580 was more effective for these parameters than atRA. Histologically, the colons were characterized by extensive ulceration with severe
mucosal inflammation, multifocal dropouts of entire crypts in
all parts of the colon, and extensive edema in vehicle-treated
mice; however, mucosal injuries were markedly attenuated
by the treatment with atRA and Am580, with slight amounts
of inflammatory infiltrates, mucosal surface erosion, and various degrees of regenerative changes without crypt depletion
(Fig. 6E). Histological grading of colitis clearly demonstrated
these differences between mice treated with RA and untreated control animals (Fig. 6F), suggesting that RA served
an important barrier-protective mechanism for the severity
of colitis by regulating epithelial integrity.
255
256
Osanai et al.
in a variety of tissues in pathological settings and demonstrated that RA depletion drives the opening of paracellular
pathways in epithelial cells, as observed in cancer cells. Our
results suggest that impaired RA signaling caused by RA
depletion leads to the disruption of functional TJs. It is thus
interesting to speculate that strategies designed to enhance
RA signals and/or blockade RA catabolism may be of therapeutic value in intestinal diseases that have increased gut
permeability. However, in short-term responses to infection,
increased permeability may have a protective function in the
gut (Grencis and Bancroft, 2004). Better understanding of
the mechanisms that regulate epithelial integrity may provide valuable insights into various intestinal disorders
(McCullough et al., 1999). It is interesting that altered gut
permeability may not be simply due to TJ-associated
gene-expression alterations but also to deregulated phosphorylation, disruption of distribution, and vacuolization of
TJ proteins (Clayburgh et al., 2004; Fasano and Shea-Donohue, 2005). In addition, TJ proteins have controversial effects
on the cells showing that overexpression of a certain type of
claudin in MDCK cells paradoxically increases paracellular
permeability (Furuse et al., 2001). This evidence suggests
Fig. 5. Cellular atRA bioavailability associates with the expression of TJ-associated genes. A, threshold cycling parameter to saturate band intensity
to the various amounts of RNA template. When we amplified cDNA reverse-transcribed from various amounts of known concentrations of RNA, a clear
correlation was observed between the amount of the template and signal intensity after PCR reaction. B, semiquantitative RT-PCR-Southern blot
analysis to examine the changes of TJ-associated genes in MDCKVec (Vec) and MDCKCYP26A1 (CYP26A1) cells after culture with or without 100 nM
atRA for up to 5 days. C, Western blot analysis of TJ proteins in MDCKVec cells after culture with or without 100 nM atRA for 3 days. ⴱ, p ⬍ 0.05 versus
cells without RA treatment.
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
necrosis factor ␣ and interferon-␥, which may be an important mechanism in gastrointestinal diseases accompanied by
barrier defects (Mankertz et al., 2000), this is the report
demonstrating that RAR␣ stimulation preferentially could
enhance the expression of TJ-associated genes such as occludin. This mechanism is partially explained by our preliminary study of the occludin promoter using approximately
2.5-kilobase upstream from the transcription initiation site,
suggesting that RAR␣ agonists stimulate occludin expression at the transcriptional level (our unpublished results).
We cannot conclude from our present study whether the RA
effect is direct on the epithelial integrity, whereas our observation may support the plausible hypothesis that strategies
that increase RA bioavailability in epithelial cells may be
promising to ameliorate intestinal morbidity in many intestinal disorders in which elevated levels of cytokines result in
gut barrier dysfunction. This is consistent with epidemiological studies showing that vitamin A supplementation significantly reduces the childhood mortality caused by the persistent diarrhea with inflammatory bowel diseases (Sommer et
al., 1986; McCullough et al., 1999).
Previous reports have shown the functional RA deficiency
Role of Retinoic Acid in Epithelial Barrier
and the in vivo study. It is therefore noteworthy to expect
that high RA bioavailability in the intestinal microenvironment preferentially affects gut epithelium, resulting in a
reduced incidence of life-threatening epithelial hyperpermeability of the gut.
It is clear that the peroxisome proliferator-activated receptor ␥ (PPAR␥) ligands such as n-3 polyunsaturated fatty
acids play an anti-inflammatory role for modulating mucosal
immunity in an animal model of inflammatory bowel disease
(Su et al., 1999; Rogler, 2006) but also that RAR activation
may compete with RXR, decreasing the activation of PPAR␥/
RXR heterodimer. Although this potential competition and
paradoxical therapeutic effects of RAR and PPAR␥ activation
after vitamin A and n-3 polyunsaturated fatty acids supplementation should be addressed in the future study, one possible explanation is that the ligand-mediated gene expression
seems to be made depending on the cell type, pharmacokinetics of the ligands, relative abundance of nuclear receptors,
extent of the competition or cross-talk among nuclear transcription factors, and the modulating role of coactivators and
corepressors on ligand-dependent transcription. In addition,
it is unlikely that only a limited amount of transcriptional
machinery and signaling is available in a given tissue.
Whether definite nuclear receptor(s) responsible for maintaining the intestinal microenvironment remains to be exam-
Fig. 6. RAR␣ stimulation attenuates the loss of intestinal barrier during TNBS-induced colitis. A, changes of body weight after the induction of TNBS
colitis. Compared with the control group (no colitis; solid line and filled symbols, n ⫽ 3 in each group), treatment group (broken line and open symbol)
with vehicle (square, n ⫽ 6) displayed a more severe clinical course with significant loss of body weight. atRA (triangle, n ⫽ 8) and Am580 (circle, n ⫽
8)-treated groups, however, did not show significant body weight loss over the course of the experiment. B, colon length relative to control animals 7
days after treatment. C, quantification of serum FITC-dextran as a measure of intestinal barrier function. D, incidence of bacterial translocation to
the spleen. E, histological sections of mouse colon treated with vehicle (DMSO), atRA, and Am580 compared with the control animals (no colitis).
Treatment with atRA and Am580 significantly protects colon from deep ulceration and severe mucosal inflammation. F, histological grading of colitis
assessed by the scoring method. ⴱ, p ⬍ 0.05 versus mice without RA treatment.
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
that the functional TJ is governed by complex regulation in
response to a wide variety of pathological insults, and future
work may reveal more detailed mechanisms explaining how
RAs maintain epithelial integrity and potentially attenuate
mucosal damage.
Vitamin A is especially critical for gut immunity, and its
deficiency results in deregulated production of cytokines such
as interferon-␥ and impaired antibody responses (Cantorna
et al., 1994). Because the intestinal mucosa is an enriched
source of cytokines that are secreted from various immunological cell types, the role of the immune system is likely to
generate an aggressive physiological response to the imbalance induced by the barrier dysfunction because intestinal
epithelial cells interface directly with various potentially
harmful antigens (Strober et al., 2003). A number of studies
have shown that RA attenuates inflammation-related epithelial damage in mouse colon (Atreya et al., 2000; Gordon et al.,
2005). Although it is impossible to conclude whether such
attenuation is due to nonspecific anti-inflammatory effects of
RA or whether it is determined by a direct protective action
of RA on mucosal epithelium, we clearly demonstrated that
RAR␣ stimulation was able to attenuate the disruption of
barrier properties of MDCK monolayer induced by the pathological insults such as oxidative stress and hyperthermia,
providing a logical connection between in vitro experiments
257
258
Osanai et al.
Acknowledgments
Various synthetic retinoids such as Am80 and Am580 were kindly
gifted from Dr. Hiroyuki Kagechika (School of Biomedical Sciences,
Tokyo Medical and Dental University, Tokyo, Japan). We thank Mr.
Kim Barrymore for help with the manuscript and the Animal Care
Facility of Sapporo Medical University School of Medicine.
References
Abu-Abed S, Dolle P, Metzger D, Beckett B, Chambon P, and Petkovich M (2001) The
retinoic acid-metabolizing enzyme, CYP26A1, is essential for normal hindbrain
patterning, vertebral identity, and development of posterior structures. Genes Dev
15:226 –240.
Asfaha S, MacNaughton WK, Appleyard CB, Chadee K, and Wallace JL (2001)
Persistent epithelial dysfunction and bacterial translocation after resolution of
intestinal inflammation. Am J Physiol 281:G635–G644.
Atreya R, Mudter J, Finotto S, Mullberg J, Jostock T, Wirtz S, Schutz M, Bartsch B,
Holtmann M, Becker C, et al. (2000) Blockade of interleukin 6 trans signaling
suppresses T-cell resistance against apoptosis in chronic intestinal inflammation:
evidence in Crohn disease and experimental colitis in vivo. Nat Med 6:583–588.
Cantorna MT, Nashold FE, and Hayes CE (1994) In vitamin A deficiency multiple
mechanisms establish a regulatory T helper cell imbalance with excess Th1 and
insufficient Th2 function. J Immunol 152:1515–1521.
Clayburgh DR, Shen L, and Turner JR (2004) A porous defense: the leaky epithelial
barrier in intestinal disease. Lab Investig 84:282–291.
de Thé H (1996) Altered retinoic acid receptors. FASEB J 10:955–960.
Dieleman LA, Palmen MJ, Akol H, Bloemena E, Pena AS, Meuwissen SG, and Van
Rees EP (1998) Chronic experimental colitis induced by dextran sulphate sodium
(DSS) is characterized by Th1 and Th2 cytokines. Clin Exp Immunol 114:385–391.
Durston AJ, Timmermans JP, Hage WJ, Hendriks HF, de Vries NJ, Heideveld M,
and Nieuwkoop PD (1989) Retinoic acid causes an anteroposterior transformation
in the developing central nervous system. Nature (Lond) 340:140 –144.
Fasano A and Shea-Donohue T (2005) Mechanisms of disease: The role of intestinal
barrier function in the pathogenesis of gastrointestinal autoimmune diseases. Nat
Clin Pract Gastroenterol Hepatol 2:416 – 422.
Furuse M, Furuse K, Sasaki H, and Tsukita S (2001) Conversion of zonulae occludentes from tight to leaky strand type by introducing claudin-2 into Madin-Darby
canine kidney I cells. J Cell Biol 153:263–272.
Gordon JN, Sabatino AD, and MacDonald TT (2005) The pathophysiologic rationale
for biological therapies in inflammatory bowel disease. Curr Opin Gastroenterol
21:431– 437.
Grencis RK and Bancroft AJ (2004) Interleukin-13: a key mediator in resistance to
gastrointestinal-dwelling nematode parasites. Clin Rev Allergy Immunol 26:51–
60.
Howe KL, Reardon C, Wang A, Nazli A, and McKay DM (2005) Transforming growth
factor-␤ regulation of epithelial tight junction proteins enhances barrier function
and blocks enterohemorrhagic Escherichia coli O157:H7-induced increased permeability. Am J Pathol 167:1587–1597.
Ishizaki T, Chiba H, Kojima T, Fujibe M, Soma T, Miyajima H, Nagasawa K, Wada
I, and Sawada N (2003) Cyclic AMP induces phosphorylation of claudin-5 immunoprecipitates and expression of claudin-5 gene in blood-brain-barrier endothelial
cells via protein kinase A-dependent and -independent pathways. Exp Cell Res
290:275–288.
Karhausen J, Furuta GT, Tomaszewski JE, Johnson RS, Colgan SP, and Haase VH
(2004) Epithelial hypoxia-inducible factor-1 is protective in murine experimental
colitis. J Clin Investig 114:1098 –1106.
Kastner P, Mark M, and Chambon P (1995) Nonsteroid nuclear receptors: What are
genetic studies telling us about their role in real life? Cell 83:859 – 869.
Kubota H, Chiba H, Takakuwa Y, Osanai M, Tobioka H, Kohama G, Mori M, and
Sawada N (2001) Retinoid X receptor alpha and retinoic acid receptor gamma
mediate expression of genes encoding tight-junction proteins and barrier function
in F9 cells during visceral endoderm differentiation. Exp Cell Res 263:163–172.
Luu L, Ramshaw H, Tahayato A, Stuart A, Jones G, White J, and Petkovich M (2001)
Regulation of retinoic acid metabolism. Adv Enzyme Regul 41:159 –175.
Mankertz J, Tavalali S, Schmitz H, Mankertz A, Riecken EO, Fromm M, and
Schulzke JD (2000) Expression from the human occludin promoter is affected by
tumor necrosis factor ␣ and interferon ␥. J Cell Sci 113:2085–2090.
Matter K, Aijaz S, Tsapara A, and Balda MS (2005) Mammalian tight junctions in
the regulation of epithelial differentiation and proliferation. Curr Opin Cell Biol
17:1– 6.
McCullough FS, Northrop-Clewes CA, and Thurnham DI (1999) The effect of vitamin
A on epithelial integrity. Proc Nutr Soc 58:289 –293.
Morris GP, Beck PL, Herridge MS, Depew WT, Szewczuk MR, and Wallace JL (1989)
Hapten-induced model of chronic inflammation and ulceration in the rat colon.
Gastroenterology 96:795– 803.
Osanai M, Murata M, Nishikiori N, Chiba H, Kojima T, and Sawada N (2006)
Epigenetic silencing of occludin promotes tumorigenic and metastatic properties of
cancer cells via modulations of unique sets of apoptosis-associated genes. Cancer
Res 66:9125–9133.
Osanai M and Petkovich M (2005) Expression of the retinoic acid-metabolizing
enzyme CYP26A1 limits programmed cell death. Mol Pharmacol 67:1808 –1817.
Prasad S, Mingrino R, Kaukinen K, Hayes KL, Powell RM, MacDonald TT, and
Collins JE (2005) Inflammatory processes have differential effects on claudins 2, 3
and 4 in colonic epithelial cells. Lab Investig 85:1139 –1162.
Prosser C, Stelwagen K, Cummins R, Guerin P, Gill N, and Milne C (2004) Reduction
in heat-induced gastrointestinal hyperpermeability in rats by bovine colostrum
and goat milk powders. J Appl Physiol 96:650 – 654.
Rogler G (2006) Significance of anti-inflammatory effects of PPARgamma agonists?
Gut 55:1067–1069.
Sommer A, Tarwotjo I, Djunaedi E, West KP Jr, Loeden AA, Tilden R, and Mele L
(1986) Impact of vitamin A supplementation on childhood mortality: a randomized
controlled community trial. Lancet 1:1169 –1173.
Strober W. Fuss IJ, Nakamura K, and Kitani A (2003) Recent advances in the
understanding of the induction and regulation of mucosal inflammation. J Gastroenterol 38:55–58.
Su CG, Wen X, Bailey ST, Jiang W, Rangwala SM, Keilbaugh SA, Flanigan A,
Murthy S, Lazar MA, and Wu GD (1999) A novel therapy for colitis utilizing
PPAR-gamma ligands to inhibit the epithelial inflammatory response. J Clin
Investig 104:383–389.
Taimi M, Helvig C, Wisniewski J, Ramshaw H, White J, Amad M, Korczak B, and
Petkovich M (2004) A novel cytochrome P450, CYP26C1, involved in metabolism of
9-cis and all-trans isomers of retinoic acid. J Biol Chem 279:77– 85.
Tobioka H, Isomura H, Kokai Y, Tokunaga Y, Yamaguchi J, and Sawada N (2004)
Occludin expression decreases with the progression of human endometrial carcinoma. Hum Pathol 35:159 –164.
Tsukita S, Furuse M, and Itoh M (2001) Multifunctional strands in tight junctions.
Nat Rev Mol Cell Biol 2:285–293.
Van Leeuwen PA, Boermeester MA, Houdijk AP, Ferwerda CC, Cuesta MA, Meyer
S, and Wesdorp RI (1994) Clinical significance of translocation. Gut 35 (1 Suppl):
S28 –S34.
White JA, Beckett-Jones B, Guo YD, Dilworth FJ, Bonasoro J, Jones G, and Petkovich M (1997) cDNA cloning of human retinoic acid-metabolizing enzyme
(hP450RAI) identifies a novel family of cytochromes P450. J Biol Chem 272:
18538 –18541.
Address correspondence to: Dr. Makoto Osanai, Department of Pathology,
Sapporo Medical University School of Medicine, South-1, West-17, Chuo-ku,
Sapporo, Japan 060-8556. E-mail: [email protected]
Downloaded from molpharm.aspetjournals.org at ASPET Journals on June 17, 2017
ined, RAR␣ is, at least in part, an important determinant for
the epithelial integrity.
Morphological examinations using immunohistochemistry
insufficiently represented the observed differences of barrier
functions between control and RA-treated MDCK cells in
animals with colitis as well. This observation supports a
concept proposed previously that barrier integrity can be
significantly compromised with no observable changes in
morphological architecture (Karhausen et al., 2004). Another
example is presented by the observation that bacterial colonization of the proximal gut leads to a leaky intestine even
without apparent mucosal damage (Asfaha et al., 2001). Alternatively, a possible explanation has emerged from the
accumulated evidence that MDCK cells display highly differentiated cellular phenotypes, expressing abundant endogenous TJ-associated proteins in cell-cell junctions (Furuse et
al., 2001).
Our present study indicates that RA provides barrier-protective elements to the epithelial cell in inflammatory mucosal injuries. Studies along this line will help us to understand
the pathogenesis of diseases characterized by the disruption
of epithelial integrity. We believe that a certain type of RA is
a promising agent for modulating the TJ function in various
diseases with gastrointestinal hyperpermeability.